Properly is like what, assembler? So lets spend 100x power on developers notebooks and 100x of their paid time and rewrite eveything in asm so it will work faster? Or what "implemented properly" means?
Dynamic languages benefit too from attention to energy-efficiency and speed in their evolution.
When PHP 7 was released, Rasmurf Lerdorf, the creator of PHP, said the performance improvements meant fewer servers, smaller memory use and reduced CPU activity - all of which equalled less power or electricity consumed.
When you consider the millions of servers in use, that additional language efficiency adds up to a substantial saving in electricity use. You can watch a segment from his presentation where he talks about this here - and the calculations he made of potential CO2 savings:
Properly in this context would mean achieving the same effect at the lowest possible overall expense of energy (including those of developer laptops) given specific hardware.
Google spends a lot of effort optimizing their software because they bear the energy cost of running it in their own data centers.
Websites do not bear the energy cost of running their code on enduser devices.
The lack of incentives turns energy waste into an externalty which means it is not factored into development.
Well, climate change doesn't care very much about developer hours. And the amount of power the dev notebooks spend is probably negligible compared to the power your production servers use. So in the context of this discussion which is explicitly about energy efficiency, I'd say this is absolutely correct.
And yes, there absolutly nonsensical stuff going on in modern software stacks, that could - in theory - be cut without any negative consequences. E.g., contemporary languages and framework seem to do a lot of stuff repeatedly at runtime that you could just as well precompute during build. However, this would require a completely different software stack than what we have now, so from our current point of view, it would be hard to change.
In a modern compiled language and without using ten layers of abstraction. It is very feasible and does not require significantly more development effort, at least not compared to the energy and material savings.
If it "would require less than a tenth of processing power" while being "very feasible and does not require significantly more development effort" I think you should get in touch with Google or Facebook with your suggestion.
You can solve two problems at once - you will lower their carbon footprint to 1/10 of their current one, and become a multi-millionaire or even billionaire. Sounds like a double win to me.
That's the problem we're talking about. It is very feasible but only if performance and energy use is something you want to focus on, while accepting the economic penalty of not prioritizing time-to-market. This is just yet another case of unchecked capitalism and free market destroying the planet. Focusing on energy and resource conservation may be more important as well as feasible, but it's not quite as economically competitive. Unless the industry as a whole decide to do what's morally right, or it is regulated and mandated by the government.
